Objective: The paper analyzes the influencing factors of hospitalization expenses in the internal medicine diagnosis groups of pulmonary malignant tumors, designs DRG grouping scheme, and provides case studies and references for the optimization of grouping scheme. Methods: The hospitalization information of patients belonging to internal medicine diagnosis groups of pulmonary malignant tumors in a Class A hospital in Luoyang City from 2019 to 2022 was collected. K-means clustering and support vector machine was used to analyze the influencing factors of hospitalization expenses, and CHAID algorithm was used to construct DRG grouping scheme. Results: Treatment methods and length of hospital stay were included in the grouping model, and 6 DRG groups were finally generated. The consistency of each DRG with the group was good, and the difference between the groups was significant, and the grouping effect was good. Conclusions: For internal medicine diagnosis groups of pulmonary malignant tumor, the grouping effect of hospitalization days is good, but it is not suitable as a grouping node. The treatment method can help to improve the DRG grouping of the subjects, but the division scheme needs to be studied.